Does the shape of a cup influence coffee taste expectations? A cross-cultural, online study

Accepted Manuscript Does the shape of a cup influence coffee taste expectations? A cross-cultural, online study George Van Doorn, Andy Woods, Carmel A. Levitan, Xiaoang Wan, Carlos Velasco, Cesar Bernal-Torres, Charles Spence PII: DOI: Reference:

S0950-3293(16)30219-1 http://dx.doi.org/10.1016/j.foodqual.2016.10.013 FQAP 3220

To appear in:

Food Quality and Preference

Received Date: Revised Date: Accepted Date:

19 July 2016 6 October 2016 30 October 2016

Please cite this article as: Van Doorn, G., Woods, A., Levitan, C.A., Wan, X., Velasco, C., Bernal-Torres, C., Spence, C., Does the shape of a cup influence coffee taste expectations? A cross-cultural, online study, Food Quality and Preference (2016), doi: http://dx.doi.org/10.1016/j.foodqual.2016.10.013

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1 Does the shape of a cup influence coffee taste expectations? A cross-cultural, online study

2

George Van Doorn1, Andy Woods2, Carmel A. Levitan3, Xiaoang Wan4,

3

Carlos Velasco2,5, Cesar Bernal-Torres6, & Charles Spence2

4 5 6

1

School of Health Sciences and Psychology, Federation University Australia, Victoria, 3842, Australia: [email protected]

7 8

2

Crossmodal Research Laboratory, Department of Experimental Psychology, University of Oxford, Oxford, UK 3

9

4

10 11 12 13

Department of Cognitive Science, Occidental College, Los Angeles, USA Department of Psychology, Tsinghua University, Beijing, China 5

6

BI Norwegian Business School, Oslo, Norway

International School of Economics and Administrative Sciences, Universidad de La Sabana, Bogota, Colombia

14 Abstract

15 16 17

We report a study designed to investigate whether shape-taste crossmodal correspondences would

18

influence consumers’ expectations concerning coffee. To that end, we conducted a cross-cultural online

19

survey with respondents (N = 309) from China, Colombia, and the United Kingdom (UK). The participants

20

had to rate eight coffee mugs on eight scales by arranging the mugs within a 1000 × 250 pixel box, placing

21

each mug so that its horizontal position matched how strongly they thought the mug matched the scale

22

presented. Amongst other findings, the results revealed that (1) the coffee was expected to be more

23

aromatic from narrower diameter mugs, (2) the coffee associated with shorter mugs was expected to be

24

both more bitter and more intense, and (3) the coffee was expected to be sweeter from wider diameter

25

mugs. An interesting cross-cultural finding was that participants from the UK expected the mugs to be

26

hotter than participants from either China or Colombia. These results add to a large and growing body of

27

research highlighting the associations between visual information and a product’s likely (or expected)

28

sensory qualities. These findings may be useful to those preparing coffee as they suggest that coffee should

29

be presented in certain mugs in order to convey a message that is congruent with the consumer’s

30

expectations.

31 32 33

Keywords: coffee, shape, mugs, taste expectations, cross-cultural, online

34 35

Introduction

36

Even before tasting, we have access to, and interpret, various pieces of sensory information concerning

37

foods and beverages (e.g., colour, orthonasal aroma, shape, and sometimes even sound and weight;

38

Prescott, 2015; Spence, 2015a; Spence & Wang, 2015). The role of this information in priming people and

39

setting their sensory and hedonic expectations

40

Blumenthal, & Blake, 2008; see also Piqueras-Fiszman & Spence, 2015, for a recent review). Shankar,

41

Levitan, and Spence (2010), for example, demonstrated that the same colour (e.g., blue) elicits different

42

expectations in different groups of people. Specifically, when a group of Taiwanese participants were

43

shown a clear plastic cup containing a blue liquid, the majority of them expected the liquid to be mint-

44

flavoured - Spence (2015b) suggests that this may be a consequence of an association with mouthwash.

45

However, when the same stimulus was shown to a group of British participants, the majority expected

46

raspberry-flavour instead. Similarly, Shermer and Levitan (2014) found that changing the colour (e.g.,

47

from red to blue) of pictures of salsa influenced participants’ expectations regarding the salsa’s spiciness.

48

However, little is known about expectations when it comes to coffee or, and similar to Shankar et al.’s

49

(2010) work, how expectations in relation to coffee might differ from one culture to the next.

50

The paucity of research exploring the influence of sensory cues on people’s expectations concerning the

51

taste/flavour of coffee is somewhat surprising, especially given Brits, for example, who are famous for

52

their fondness for tea, consume an estimated 70 million cups of coffee in cafés, restaurants, and other

53

outlets each and every day (Howie, 2012)2. Such figures hint at the ubiquity of coffee in many countries

54

(see P. J. W., & D. H., 2013) and, given the economic incentive to keep consumers drinking coffee, café

55

owners, restaurateurs, crockery designers and manufacturers ought, presumably, to be interested in

56

anything that helps enhance the perception of the taste qualities, the enjoyment, or the overall coffee

57

drinking experience for their clientele (cf. Van Doorn, Wuillemin, & Spence, 2014).

58

Shape-taste associations

59

Shape undoubtedly influences consumer behaviour (see Spence, 2012, for a review), and any shapes that

60

are present on, or near, a food or beverage can be used by consumers to assess the likely qualities of that

61

foodstuff. In general, people prefer rounded shapes (e.g., circles) to more angular shapes (e.g., triangles or

62

stars; Bar & Neta, 2006; Gómez-Puerto, Munar, & Nadal, 2015; Silvia & Barona, 2009). Cheskin’s (1957)

63

oft-cited research drew attention to the impact of shapes on people’s perception of different products.

64

Cheskin placed identical products (e.g., crackers) in two different packages, one adorned with triangles, the

65

other with circles. The participants’ task was to state which product they preferred. Eighty-percent of

66

participants reported a preference for the product from the package adorned with circles; often suggesting,

1

has been well-established (Yeomans, Chambers,

1

Consistent with Olson and Dover (1976), an expectation is defined here as “the perceived likelihood that a product possesses a certain characteristic or attribute” (p. 169). 2

This figure includes the cups of coffee drunk at home and in other locations (e.g., staff tea rooms); approximately 70% of which are instant coffee.

67

when quizzed, that this was of better quality. Westerman et al. (2012) obtained similar results in relation to

68

people’s preference for rounded shapes on, and rounded contours of, product packages.

69

Shape also seems to have a role in the experience when drinking a beverage (see Hanson-Vaux, Crisinel, &

70

Spence, 2013). Demonstrating a tangible impact of shape on drinking, Wansink and van Ittersum (2003,

71

2005) found that both children and adults pour around 20-30% more of a drink (e.g., juice) into short/wide

72

glasses relative to tall/thin glasses. However, participants believed the opposite to be true. These authors

73

related this finding to Piaget’s conservation task. Specifically, adults fail the task because it appears as

74

though they believe that tall/thin containers hold more fluid than short/wide containers, and thus they pour

75

less fluid into tall/thin containers.

76

Although associations between shape and taste have been explored in a range of food and beverage

77

products, the correspondence between shape and expectations related to the taste of coffee remain

78

unknown. Coffee is an interesting candidate for research because of its consistent, bitter character and the

79

different bitter/sweet combinations that arise through bean selection, type of roasting of the beans, type of

80

milk used (e.g., full fat), and whether or not sugar is added. According to Spence (2012), coffee is likely to

81

be another product where shape-taste associations exist. The suggestion being that many coffee company

82

logos are rounded in shape (e.g., New York Coffee Company, Costa Coffee, Starbucks Coffee), and that

83

this might be used to suggest to customers that their coffee is not overly bitter (see also Batra, Seifert, &

84

Brei, 2015; Zhang, Feick, & Price, 2006). However, it is important to note that this claim has yet to be

85

substantiated, and Cheskin’s (1957) early ideas (i.e., the ability of the shapes used on product packaging to

86

affect people’s product expectations) have yet to be applied to the coffee category. This research project

87

addresses this salient gap in the literature. Specifically, and given that, in a restaurant setting, a coffee’s

88

package is often the mug or cup in which it is served, we sought to investigate shape-flavour associations

89

in relation to coffee expectations.

90

Cross-cultural research

91

Interestingly, Bremner et al. (2013) reported that the Himba tribe of Kaokoland in rural Namibia did not

92

show the ‘usual’ (i.e., Western) associations between angular and rounded shapes and the tastes and oral-

93

somatosensory properties of beverages. It was assumed that the Himba have been unable to accumulate the

94

‘usual’ associations through experience because they have not been exposed to written language,

95

supermarkets, or advertising. Bremner et al. found that the Himba did not match still water with an organic,

96

amoeba-like shape, nor did they pair sparkling (i.e., carbonated) water with an angular, star-like shape.

97

Additionally, they also matched chocolates varying in cocoa content in a manner opposite to that of their

98

Western counterparts (i.e., Westerners match chocolate high in cocoa to angular, star-like shapes due to the

99

increased bitterness). That said, Ngo et al. (2013) have observed consistent crossmodal correspondences

100

across cultures. Specifically, they demonstrated that British and Colombian participants associated sweet

101

fruit juices with round shapes and sour fruit juices with angular shapes (see also Salgado-Montejo et al.,

102

2015; Wan et al., 2014). Bremner et al.’s (2013) findings, and the work of others (e.g., Williams & Bargh,

103

2008), show that at least some of the associations between shapes and the tastes, flavours, aromas, and

104

oral-somatosensory attributes of food and beverages are likely learned. That said, it is possible that

105

participants matched stimuli as a function of stimulus valence, which might differ across cultures (see

106

Velasco, Woods, Petit, Cheok, & Spence, 2016). For example, the Himba might find both chocolate high

107

in cocoa and rounded forms appealing, and thus match them.

108

Aims and hypotheses

109

In the study reported here, we explored the impact of the shape of coffee mugs on people’s expectations of

110

the coffee. Most studies on taste/shape associations have focused on the curvilinearity of shapes. However,

111

other shape features (in particular those that affect visual preference) may influence taste/shape

112

associations (as shown by Salgado-Montejo et al., 2015, for symmetry; Deroy & Valentin, 2011, for

113

thinness). Further, and similar to Piqueras-Fiszman, Alcaide, Roura, and Spence (2012), we wanted to

114

explore the influence of the shape of the container the beverage is served in. For those reasons we explored

115

some of the attributes that are typically varied in coffee cups, namely the ‘height’ of the mug (tall, short),

116

the ‘diameter’ of the mug (wide, narrow), and the ‘thickness’ of the rim (thick, thin). It should be noted

117

that factors other than shape can influence expectations as well. For example, the cup in which the coffee

118

is served may affect us as a function of our perception of the general properties of the cup (i.e., cheap vs.

119

expensive [Piqueras-Fiszman, Harrar, Alcaide, & Spence, 2011], flimsy vs. strong [Krishna & Morrin,

120

2008]). Here, we explore these issues too.

121

In the remainder of this section, the hypotheses will be discussed according to the type of expectation

122

measured. Specifically, ‘bitterness’ and ‘sweetness’ measure expectations relating to the taste of coffee,

123

while ‘aroma’, ‘energy’, ‘temperature’, and ‘intensity’ measure expectations concerning the

124

properties/qualities of coffee. Finally, ‘liking’ and ‘willingness-to-pay’ measure people’s expectations

125

concerning themselves.

126

Taste Expectations

127

It was thought that if expectations are affected by a mug’s attributes (e.g., height), a coffee’s properties

128

(e.g., bitterness) should be rated more favourably when associated with a particular change in that

129

dimension. For example, it is common in several countries to serve more concentrated coffees (e.g.,

130

espresso, macchiato) in smaller cups and, as such, we expected people to rate these mugs as containing

131

coffees that were more bitter.

132

Expectations regarding the coffee’s properties

133

It is possible that different cup diameters influence expected aroma intensity. Cliff (2001) suggested that

134

larger openings allow aromas to escape prior to evaluation, and the same logic could be applied here. That

135

said, Spence (2011, 2016) suggested that a small-diameter glass reduces the surface area of the contents

136

available for diffusion, and thus fewer odour molecules are released from the liquid. Given these

137

conflicting findings, we thought it most appropriate to hypothesise that ‘cup diameter’ would not influence

138

the expected aroma of coffee.

139

Expectations relating to the individual

140

It was hypothesised that increases in ‘cup height’ and ‘cup diameter’ would be associated with an increase

141

in the amount a person was willing-to-pay for the coffee, due to the expectation that there will be more

142

coffee in these cups. Importantly though, and consistent with Wansink and van Ittersum (2003, 2005), it

143

may be that people pay more attention to one dimension of the cup (e.g., height) than another (e.g., width).

144

If this is true, and Wansink and van Ittersum are correct, it was thought that people might expect that

145

tall/thin mugs hold more coffee relative to short/wide mugs. As such, people would be willing-to-pay more

146

for coffee from these types of mugs.

147

Consistent with Harrar and Spence (2013), it was thought that the thickness of the mugs would influence

148

expected attributes of the coffee. This thought is based on the fact that thicker objects (usually) weigh

149

more than thinner objects. Harrar and Spence found that yoghurt was perceived of as being more expensive

150

when it was tasted from a lighter plastic spoon, relative to an artificially-weighted spoon. As such, we

151

hypothesised that the coffee associated with thin-walled mugs, which one assumes are expected to be

152

relatively lighter, would be deemed more expensive than the coffee associated with mugs with thicker

153

walls. However, it could be argued that, in Harrar and Spence’s work, there is a contrast between the

154

weight of the spoon and the perceived thickness/creaminess (and thus expensiveness) of the yoghurt. In the

155

study presented here, though, there was no real coffee, so there is no contrast. Consequently, it might be

156

that people expect higher quality coffee to come in thicker cups.

157

Method

158

Participants

159

Three hundred and nine participants took part in the study. One hundred and three volunteers (46 women)

160

aged between 17 and 29 years were from China (Mage = 21.50 years, SDage = 8.07 years). Ninety-seven

161

volunteers (56 females) aged between 18 and 69 years were from Colombia (Mage = 29.19 years, SDage =

162

14.21 years). Finally, 105 participants (52 females) aged between 16 and 60 years were from the UK (Mage

163

= 34.10 years, SDage = 11.05 years).

164

The Chinese participants were undergraduate or graduate students from Tsinghua University, Beijing,

165

China. For their participation, volunteers received either course credit in order to fulfil the requirements of

166

an introductory psychology course that they were enrolled in, or were compensated ¥12.5 CNY. The

167

experiment was approved by the ethics committee at the Psychology Department of Tsinghua University,

168

and conformed to the ethical standards for conducting research established by the American Psychological

169

Association. The Colombian participants were recruited from a database of participants created at the

170

International School of Economic and Administrative Sciences at Universidad de La Sabana, Bogota,

171

Colombia, and took part in the experiment voluntarily. The UK participants were recruited from Prolific

172

Academic to take part in the study in return for a payment of 1.00 UK pound. By means of Prolific

173

Academic’s ‘filter’ feature, only those participants who reported having been born in the UK were allowed

174

to take part in the study. The study was reviewed and approved by the Central University Research Ethics

175

Committee at Oxford University and was carried out in accordance with the World Medical Association

176

(WMA, 2013) Helsinki Declaration. All participants provided informed consent prior to taking part in the

177

study.

178 179

Stimuli

180

Given that the experiment was conducted online, the apparatus varied by participant. Nevertheless, the

181

experiment utilized ‘full screen’ mode (i.e., utilizing the entirety of the participant’s monitor), and took

182

place within a 1024 × 768 pixel box in the centre of the screen (see Figure 1), irrespective of the size of the

183

participant’s monitor. The experiment was conducted online using the Adobe Flash-based version of

184

Xperiment (http://www.xperiment.mobi).

185 186

Figure 1. The pictures used in the survey.

187 188

Design

189

A mixed-factorial design was used that included a between-participant factor (country of origin: China,

190

Colombia, or the UK) and the within-participants factors of the ‘height of cup’ (tall, short), the ‘cup

191

diameter’ (wide, narrow), and the ‘thickness of rim’ (thick, thin). The dependent variables are defined in

192

Table 1. Note that due to human error whilst scripting the study, participants from the UK were asked to

193

specify how much they would pay for drinks in terms of US dollars, not UK pounds.

194 195

196 197 198 199 200 201 202

Table 1. The dependent variables, the question asked to assess each, and the anchors used to define the scale participants had to place the mugs along (the anchors were always placed on the far left and right of the scale; in the case of ‘Willingness-to-pay’ though, the additional anchors were evenly spaced between the far left and far right anchor). Dependent variable

Question asked

Scale anchors (left to right)

Aroma

Please arrange these mugs of coffee in order of how strong smelling you would expect the coffee from each to be

Not aromatic at all; Very strongly aromatic

Bitter

Please arrange these cups of coffee in order of how bitter you would expect each to taste

Not bitter at all; Very bitter

Energy

Please arrange these mugs in order of how energising you think the coffee in each would be

Not at all energising; Very energising

Temperature

Please arrange these cups in order of how hot you would expect the coffee from each to be

Body temperature; Too hot to hold

Intensity

Please arrange these mugs of coffee in order of how intense you would expect coffee from each to taste

Not intense at intense

Liking

Please arrange these mugs of coffee in order of how much you expect to like the coffee from each

Greatest imaginable dislike; Greatest imaginable like

Sweetness

Please arrange these mugs of coffee in order of how sweet you would expect coffee from each to taste

Not sweet at all; Very sweet

Willingnessto-pay

Please arrange these mugs of coffee in order of how much money you would be willing to pay for a cup of coffee in each

English: 0 - 10 US dollars

all; Very

Chinese: 0 - 45 Chinese Yen Colombia: 0 - 31000 $Pesos

203 204

Procedure

205

A screen shot of the task is shown in Figure 1. The participants had to arrange the mugs within a 1000 ×

206

250 pixel box, placing each mug so that its horizontal position matched how strongly they thought each

207

mug matched the scale presented (e.g., in Figure 1, the participant is being asked to arrange the mugs

208

according to how hot they think coffee presented in each will be). Mugs could be placed so that they

209

overlapped (with the most recently moved placed on top of mugs moved earlier). Parenthetically, the mugs

210

we showed to participants did not have coffee in them and we (deliberately) did not specify whether there

211

was the same amount of coffee in each cup. As such, each participant may have had a different idea with

212

regards to the ‘amount’.

213

After placing all eight mugs, the participant could proceed to the next trial by pressing the space bar or

214

clicking the ‘next’ button (there was a 100ms pause between trials). On each of the eight trials, a different

215

scale was presented. The original starting positions for the mugs were arranged randomly in a 1000 × 269

216

pixel area above the box (if a mug’s random placement overlapped with another mug, a new random

217

placement was generated; this was repeated up to 100 times, after which the mug was placed in the

218

position that, out of the prior 100 attempts, least overlapped existing mugs). Trial order was randomised

219

between participants 3 . The participants took an average of 650 seconds to complete the study. After

220

completing all the trials participants were debriefed as to the nature of the study. This kind of task has been

221

used successfully in several recent studies (e.g., Velasco, Woods, Hyndman, & Spence, 2015).

222

Analyses

223

Eight mixed-factorial ANOVAs, subjected to Holm-Bonferroni corrections, were conducted that were

224

identical in terms of design except for their dependent variable (Aroma, Bitterness, Energy, Temperature,

225

Intensity, Liking, Sweetness, and Willingness-to-pay); the dependent variable was the position on the x-

226

axis of the centre of the images of the coffee mugs, relative to the size of the box within which the mugs

227

were placed - percentage position values were used. In relation to the Holm-Bonferroni corrections, there

228

were 15 main effects and interactions per ANOVA, so the most stringent critical p-value used was 0.05 /

229

(15 x 8) = 0.00042; critical p-values and statistics are detailed in Appendix 1. Contrary to popular opinion,

230

ANOVA does not control for Type 1 error (see Lakens, 2016). Each ANOVA consisted of the between-

231

participant factor of ‘country of origin’ (China vs. Colombia vs. UK), and the repeated-measures factors of

232

‘height of cup’ (tall vs. short), ‘cup diameter’ (narrow vs. wide), and ‘thickness of rim’ (thick vs. thin). The

233

full report of these analyses is given in Appendix 1.

234

Results

235

Data screening

236

Outliers were screened, and corrected separately, for each country (values exceeding 3 x SD +/- mean were

237

replaced with the next most extreme, but non-outlying, value). Eleven out of 6720 data points were

238

corrected in this fashion for UK data, and 11/6208 for Colombian data (none of the 6592 Chinese data

239

points were outliers).

240

Taste Expectations

241

Bitterness

242

Although the three-way interaction between ‘thickness of rim’, ‘height of cup’, and ‘country of origin’ was

243

significant [F(2, 302) = 9.32, p < .001, η2p = .06], inspection of the data (see Figure 2) indicates that

244

‘height of cup’ was more impactful than ‘thickness of rim’ and/or ‘country of origin’. This is supported by

245

the fact that the only main effect, from these three factors, that reached statistical significance was ‘height

246

of cup’ [F(2, 302) = 69.04, p < .001, η2p = .19]. Here, the coffee associated with short mugs (M = 58.62; CI

3

Please contact Andy Woods ([email protected]) for the script for the Cantonese and Spanish versions of the text used in the study.

247

[56.76, 60.48]) was expected to be more bitter than the coffee associated with taller mugs (M = 45.34; CI

248

[43.46, 47.21]). There was also a significant main effect of ‘cup diameter’ [F(1, 302) = 137.56, p < .001,

249

η2p = .31], with the coffee associated with narrower diameter mugs (M = 64.07; CI [61.69, 66.46]) thought

250

to be more bitter than the coffee associated with wider diameter mugs (M = 39.89; CI [37.74, 42.03]).

251

Table 2 presents a summary of all the significant main effects.

252 100 90 80

Bitterness

70 60 50 40 30 20 10

Tall

Short

Tall

Thick

UK

Colombia

China

UK

Colombia

China

UK

Colombia

China

UK

Colombia

China

0

Short Thin

Cup attributes

253 254 255

Figure 2. The interaction between ‘thickness of rim’, ‘height of cup’ and ‘country of origin’ for Bitterness (error bars here and henceforth represent the 95% CI around the mean).

256 257

Table 2. A summary of the significant main effects. Expectations

Taste

Quality

Subjective ratings

DV

Main effects Height of cup

Diameter of cup

Thickness of rim

Country of origin

Bitter

√

√

-

-

Sweetness

-

√

-

-

Aroma

√

√

-

-

Energy

-

-

-

-

Temperature

-

-

-

√

Intensity

√

√

-

-

Liking

-

-

-

-

Willingness-to-pay

258

√

√

-

-

Note: √ denotes a significant main effect

259 260 261 262 263

Sweetness

264

The main effect of ‘cup diameter’ achieved significance [F(1, 302) = 33.55, p < .001, η2 p = .10), with the

265

coffee from mugs with a wider diameter (M = 55.38; CI [53.05, 57.71]) expected to be sweeter than coffee

266

from mugs with a narrower diameter (M = 42.40; CI [39.75, 45.05]).

267

Expectations regarding the coffee’s properties

268

Aroma

269

The main effects of ‘cup diameter’ [F(1, 302) = 13.78, p < .001, η2 p = .04] and ‘height of cup’ [F(1, 302) =

270

45.73, p < .001, η2p = .13] exerted a significant influence on participants’ ratings of expected aroma. In

271

terms of ‘cup diameter’, the coffee associated with narrower diameter mugs (M = 59.32; CI [56.64, 62.01])

272

was expected to be more aromatic than the coffee associated with wider diameter mugs (M = 50.77; CI

273

[48.42, 53.12]). In relation to ‘height of cup’, the coffee from short mugs (M = 60.47; CI [58.50, 62.45])

274

was thought to be more aromatic than was the coffee from taller mugs (M = 49.62; CI [47.74, 51.50]).

275

Energy

276

There were no significant main effects or interactions (see Appendix 1).

277

Intensity

278

The main effects of ‘cup diameter’ [F(1, 302) = 110.67, p < .001, η2p = .27] and ‘height of cup’ [F(1, 302)

279

= 81.51, p < .001, η2p = .21] were significant. The coffee associated with narrower diameter mugs (M =

280

64.61; CI [62.09, 67.12]) was expected to be more intense than that associated with wider diameter mugs

281

(M = 42.12; CI [40.02, 44.22]). Likewise, coffee in short mugs (M = 60.56; CI [58.66, 62.46]) was

282

expected to be more intense than coffee from tall mugs (M = 46.17; CI [44.39, 47.95]).

283

Temperature.

284

The only main effect that achieved statistical significance here was ‘country of origin’ [F(2, 302) = 12.89,

285

p < .001, η2p = .08], with UK participants expecting the mugs to be hotter (M = 55.50; CI [53.61, 57.39])

286

than participants from either China (M = 50.14; CI [48.23, 52.04]) or Colombia (M = 48.96; CI [47.00,

287

50.93]).

288

Expectations relating to the individual

289

Liking

290

The interaction between ‘height of cup’ and ‘country of origin’ achieved significance [F(2, 302) = 9.90, p

291

< .001], with a medium effect size (η2p = .06). Figure 3 shows that the interaction was largely driven by

292

Chinese participants liking coffee from short mugs (M = 58.90; CI [55.28, 62.51]) relative to taller mugs

293

(M = 42.94; CI [39.61, 46.26]). Confidence intervals revealed that Colombians’ liking of coffee from short

294

[50.25, 57.71] and tall mugs [49.58, 56.44] and UK participants’ preference for coffee from short [50.82,

295

57.99] and tall mugs [51.04, 57.63] overlapped – but were greater than the Chinese participants liking for

296

coffee from tall mugs. 100 90 80

Liking

70 60 50 40 30 20 10 0 Tall

Short China

Tall

Short

Tall

Colombia

Short UK

Cup attributes

297 298

Figure 3. The interaction between ‘height of cup’ and ‘country of origin’ for Liking.

299

Willingness-to-pay

300

Chinese Yen (6.214 CNY = 1 USD) and Colombian Peso (2382 COP = 1 USD) were converted to US

301

dollars using the currency exchange rate midway through testing (20th January, 2016, via

302

http://www.exchangerates.org.uk/). We were interested in the relative changes as a function of our

303

experimental conditions and although the amounts may represent something different in each country, they

304

nevertheless provide us with the relative changes, in terms of the manipulation of interest. Given that the

305

study was conducted over a 6 month period, and given the degree of variation of the exchange of these

306

currencies (which, even if the relative value of the currencies remained stable, could have many possible

307

explanations), we decided to focus more on within country variation in the Discussion as opposed to

308

variation across countries.

309

There was a significant interaction (see Figure 4) between ‘cup diameter’ and ‘country of origin’ [F(2, 302)

310

= 28.71, p < .001, η2p = .16]. Whilst both Colombians (M = 5.38; CI [4.81, 5.95]) and participants from the

311

UK (M = 4.68; CI [4.38, 4.99]) rated coffee from wider diameter mugs as being more expensive than

312

coffee from mugs with a narrower diameter (Colombians: M = 2.93; CI [2.65, 3.21]; UK: M = 3.51; CI

313

[3.24, 3.78]), Colombians reported that they were willing-to-pay less for coffee from smaller diameter

314

mugs than were participants from the UK. The amount Chinese participants were willing-to-pay for coffee

315

did not depend on the diameter of the cup (i.e., wide diameter: M = 3.55; CI [3.31, 3.80]; narrow diameter

316

M = 3.51; CI [3.20, 3.81]).

317 10 9 8

Willingness to pay ($)

7 6 5 4 3 2 1 0

Wide diameter

318 319

Narrow diameter

China

Wide diameter

Narrow diameter

Colombia

Wide diameter

Narrow diameter UK

Figure 4. The interaction between ‘cup diameter’ and ‘country of origin’ for the Willingness-to-pay DV.

320 321

The interaction (see Figure 5) between ‘height of cup’ and ‘country of origin’ also achieved significance

322

[F(2, 302) = 20.04, p < .001], with a medium effect size (η2p = .12). The interaction is almost identical to

323

the previous interaction (see Figure 4). Specifically, both Colombians (M = 4.96; CI [4.48, 5.44]) and

324

UK participants (M = 4.53; CI [4.21, 4.79]) were willing-to-pay more for coffee from tall mugs than they

325

were for coffee from short mugs (Colombians: M = 3.35; CI [3.02, 3.67]; UK: M = 3.66; CI [3.37, 3.96]),

326

whereas the amount Chinese participants were willing-to-pay did not depend on the height of the mug (i.e.,

327

tall: M = 3.54; CI [3.29, 3.79]; short: M = 3.52; CI [3.28, 3.76]).

328

10 9

Willingness to pay ($)

8 7 6 5 4 3 2 1 0

Tall China

329 330

Short

Tall

Short Colombia

Tall

Short UK

Figure 5. The interaction between ‘height of cup’ and ‘country of origin’ for the amount one was Willing-to-pay.

331 332

‘Cup diameter’ and ‘height of cup’ also interacted [F(1, 302) = 12.83, p < .001, η2p = .04]. People were

333

willing-to-pay the most for tall/wide cups (M = 5.06; CI [4.75, 5.38]), followed by short/wide cups (M =

334

3.98; CI [3.76, 4.21]) and tall/narrow mugs (M = 3.60; CI [3.41, 3.79]), which did not differ from one

335

another, and, finally, short/narrow mugs (M = 3.05; CI [2.83, 3.26]).

336

The main effects of ‘cup diameter’ [F(1, 302) = 90.62, p < .001, η2 p = .23] and ‘height of cup’ [F(1, 302) =

337

66.10, p < .001, η2p = .18] exerted a significant influence on the amount participants’ were willing-to-pay.

338

Unsurprisingly, and in relation to ‘cup diameter’, people were willing-to-pay more for coffee from mugs

339

with a wider diameter (M = 4.57; CI [4.37, 4.78]) than they were for coffee from narrower diameter mugs

340

(M = 3.53; CI [3.35, 3.72]). As for ‘height of cup’, people were willing-to-pay less for coffee from short

341

mugs (M = 3.70; CI [3.53, 3.87]) than they were for coffee from taller mugs (M = 4.41; CI [4.22, 4.60]).

342

Discussion

343

The main issue explored in this study was whether expectations about coffee are influenced by changes in

344

the shape of the mug. The results revealed that ‘cup diameter’ and ‘cup height’ influenced the expected

345

aroma, bitterness, intensity, and amount a participant was willing-to-pay; ‘cup diameter’ also influenced

346

the expected sweetness. An interesting cross-cultural finding was that participants from the UK expected

347

the mugs to be hotter than participants from either China or Colombia. In contrast to Harrar and Spence’s

348

(2013) finding relating to the weight of spoons, the weight (which was assumed to be associated with

349

‘thickness’) of the mugs did not influence expected attributes of the coffee – this seems odd given that tea

350

drinkers would presumably consider ‘cup thickness’ an important issue (consider, for example, the thin lip

351

of a bone China cup). Harrar and Spence found that yoghurt was thought to be more expensive when it was

352

tasted from a lighter plastic spoon, relative to an artificially-weighted spoon. As such, we initially thought

353

that coffee associated with thin-walled mugs, which one assumes are expected to be relatively lighter,

354

would be considered more expensive than the coffee associated with mugs with thicker walls. However,

355

some literature (e.g., Piqueras-Fiszman et al., 2011; Piqueras-Fiszman & Spence, 2012) suggests that the

356

coffee associated with thick-walled mugs, which one assumes are expected to be relatively heavier, would

357

be deemed more expensive than the coffee associated with mugs with thinner walls. Further, in Harrar and

358

Spence’s work there was a contrast between the weight of the spoon and the perceived

359

thickness/creaminess (and thus expensiveness) of the yoghurt. Consequently, it might be that people expect

360

higher quality coffee to come in thicker cups. Neither of these hypotheses were supported, which may be a

361

consequence of the fact that our task measured expectations, whereas Harrar and Spence (2013) tested

362

perceptions. It might also be true that, because we used conservative Holm-Bonferroni corrections, effects

363

that achieved significance in previous work did not do so here. However, the null finding might be an

364

artefact of the stimuli we used. It is possible that participants had difficulty distinguishing the two variable

365

levels (i.e., thick walls vs. thin walls), and thus provided similar responses regardless of the ‘thickness of

366

rim’.

367

Taste Expectations

368

Bitterness

369

The coffee associated with short mugs was expected to be more bitter than the coffee associated with taller

370

mugs. A seemingly logical interpretation of this finding is that people (from several cultures) expect the

371

ratio of coffee to milk (or water) in the shorter mugs to be greater than they expect the ratio to be in taller

372

mugs, and thus expect the coffee in shorter mugs to be more bitter. Similarly, perhaps it is that people

373

expect certain types of coffees to be served in smaller cups. For example, in the UK and Australia, it is

374

common for “strong” coffees (think espresso, macchiato) to be served in very small cups. At this point, it

375

is worth considering that features such as ‘cup height’ may be matched to specific taste attributes. Here, we

376

are dealing with the specific semantic context of ‘coffee’, and in that sense people may filter information

377

as a function of their ‘experience’ with coffee (see Bohrn, Nabecker, & Carbon, 2008; Carbon, 2010 for

378

similar arguments in relation to shape curvature preference).

379

This same logic can be applied to the finding that ‘cup diameter’ was significant. Specifically, the coffee

380

associated with narrow-diameter mugs was thought to be more bitter than the coffee associated with wide-

381

diameter mugs. Again, and holding mug height constant, it may be that people expect the ratio of coffee to

382

milk (or water) in the narrower mugs to be greater than it is in wider mugs, and thus expect the coffee in

383

narrower mugs to be more bitter.

384

Sweetness

385

The main effect of ‘cup diameter’ achieved statistical significance, with the coffee from mugs having a

386

wider diameter expected to be sweeter than coffee presented in mugs having a narrower diameter. This

387

might be the inverse of the “bitterness” finding. Specifically, the coffee associated with mugs with a

388

narrower diameter was thought to be less sweet (or more bitter) than the coffee associated with mugs of a

389

wider diameter. Again, one possibility here is that people expect the ratio of coffee to milk (or water) in the

390

wider diameter mugs to be less than it is in narrower mugs, and thus expect the drink to be less bitter (or

391

sweeter).

392

Expectations regarding the coffee’s properties

393

Aroma

394

To reiterate, the main effects of ‘cup diameter’ and ‘height of cup’ exerted a significant influence on

395

participants’ ratings of the expected aroma. Although it is difficult to disentangle the important factors in

396

the work of Cliff (2001), the results presented here seem to be (somewhat) consistent with her findings in

397

relation to wine. Specifically, we found that the coffee associated with smaller diameter mugs was thought

398

to be more aromatic than the coffee associated with larger diameter mugs. Cliff found that wine glasses

399

with large bowl diameters but small openings had the highest aroma intensities, regardless of the type of

400

wine sampled. Cliff suggested that larger openings allow aromas to escape prior to evaluation, and the

401

same logic could be applied here. However, Spence (2011) suggested that a small-diameter glass reduces

402

the surface area of the contents that is available for diffusion, and thus fewer odour molecules are released

403

from the liquid. Coffee might be an interesting case where expectations and perceptions differ.

404

In relation to ‘height of cup’, the coffee from short mugs was thought to be more aromatic than that from

405

taller mugs. Although speculative, this finding (and the finding regarding ‘cup diameter’) might, again, be

406

related to bitterness and the idea that people filter information as a function of their experiences. It might

407

also relate to the work of Jeon, Lee, and Kim (2014) who highlight the importance of expectations. Jeon

408

and colleagues showed that people expect soup to be presented in certain type of bowls, and this

409

expectation can influence its perceived saltiness. The same logic could be applied here in that it is common

410

in several countries to serve more concentrated coffees in smaller cups and, as such, people might expect

411

coffees presented in these mugs to be more aromatic.

412

Energy

413

None of the main effects or interactions achieved significance. As such, the coffee associated with certain

414

mug types was not deemed more energizing than the coffee associated with any other mug type.

415

Supporting the null hypothesis here is interesting because one might assume that there is a correlation

416

between ‘energy’ and ‘volume’. Consider, for example, energy drinks: A relatively uncontroversial

417

assumption would be that people expect larger volumes of energy drink to be more energizing than smaller

418

volumes. It is, therefore, somewhat surprising that people do not expect larger volumes of a similarly

419

caffeinated beverage (i.e., coffee) to be more energizing. A tentative explanation here is that the coffee

420

category might be somewhat unique. That is, people understand that smaller coffees (e.g., espresso) are

421

usually quite strong, and that larger coffees (e.g., lattes) often have an equivalent amount of coffee in them,

422

but are topped-up with milk and foam.

423

Temperature

424

There was a main effect of ‘country of origin’. Here, participants from the UK expected the mugs to be

425

hotter than did the participants from either China or Colombia. An interesting, yet speculative, idea here is

426

that people from the UK expect coffees to be warmer because the climate (13.5oC) there is, on average,

427

colder than it is in Bogota (Colombia: 18.0o C) and Beijing (China: 17.8oC). This proposition, obviously,

428

requires further testing.

429

Intensity

430

The main effects of ‘cup diameter’ and ‘height of cup’ were significant. The coffee associated with the

431

narrower diameter cups was expected to be more intense than that associated with wider mugs. Likewise,

432

coffee in short mugs was expected to be more intense than that from tall mugs. Interestingly, these findings

433

mimic those for bitterness. Consistent with an argument made by Van Doorn, Wuillemin, and Spence

434

(2014), consumers appear to blur the distinction between ‘intensity’ and ‘bitterness’. Dijksterhuis (1998)

435

has suggested that because of the use of the word ‘strong’ in coffee advertising, consumers often confuse a

436

coffee’s strength or intensity with its ‘bitterness’ – the finding here that intensity ratings mirror bitterness

437

ratings would support such a view.

438

Expectations relating to the individual

439

Liking

440

The interaction between ‘height of cup’ and ‘country of origin’ was significant, and driven largely by

441

Chinese participants’ preference for coffee in short mugs. Colombians and participants from the UK

442

showed no preference for coffee from either short or tall mugs. However, both groups rated the coffee in

443

these mugs as being more likeable than was Chinese participants rating of the coffee in tall mugs (see

444

Figure 3). A possible explanation for this findings is that participants might simply be responding as a

445

function of the ‘regularities’ found in coffee drinking experiences, over-and-above any crossmodal feature

446

matching. More work is needed to clarify this issue.

447

Willingness-to-pay

448

There was a significant interaction between ‘diameter of cup’ and ‘country of origin’. Whilst both

449

Colombian and UK participants were willing-to-pay more for coffee from mugs having a wider (as

450

compared to a narrower) diameter, the Chinese participants failed to differentiate between narrow and wide

451

diameter mugs with respect to the amount they were willing-to-pay. This seems like an odd finding but,

452

perhaps, is a consequence of the fact that coffee is still not a common beverage in China. That is,

453

Colombians and those from the UK hold an expectation that a greater volume of coffee (as one would get

454

in a wider diameter mug) would cost more but, due to their lack of familiarity with coffee, Chinese

455

participants did not necessarily expect to pay more for a slightly larger quantity. The interaction between

456

‘height of cup’ and ‘country of origin’ mimics the interaction between ‘diameter of cup’ and ‘country of

457

origin’ and the same explanation seems applicable. That said, as Chinese participants were younger than

458

those from either Colombia or the UK, willingness-to-pay might be influenced by (possible) differences in

459

coffee consumption patterns and income, regardless of the shape of mug. Further investigation is required.

460

There was a significant interaction between ‘cup diameter’ and ‘height of cup’ that demonstrated that

461

participants were willing-to-pay the most for tall/wide cups, and the least for short/narrow mugs.

462

Unsurprisingly, this finding suggests that willingness-to-pay is better explained by the perceived volume of

463

the coffee, as opposed to the individual factors of ‘height of cup’ and ‘cup diameter’. This interpretation is

464

supported by the significant main effects of ‘cup diameter’ and ‘height of cup’ – where people were

465

willing-to-pay less for smaller cups of coffee relative to larger cups of coffee. Interestingly, the findings

466

do not seem to support those of Wansink and van Ittersum (2003, 2005). In the present study, the

467

willingness-to-pay CIs for the short/wide mug overlap those of from the tall/narrow mug. As such, one

468

could draw the conclusion that adults expected these mug types to hold an equivalent amount of coffee.

469

Limitations

470

There are several issues that may have influenced our results and should be considered. The first, as raised

471

by a reviewer, was that the participants from the different countries had different mean ages and it could be

472

the case that coffee consumption varies as a function of age. A further two differences were that whilst

473

participants from China and the Colombia were students recruited through their universities, those from the

474

UK were recruited through the online recruitment panel www.prolificacademic.co.uk. Further, participants

475

recruited in Colombia did not receive monetary compensation for taking part. It is less clear if these factors

476

would have influenced our results, nevertheless, it is worth outlining these as potential confounds to avoid

477

in future studies related to ours.

478

Conclusions

479

The results of the survey reported here demonstrate that the shape of the mug influenced people’s

480

expectations of the taste and qualities of coffee that would be served in such a mug. Shape, or more likely

481

‘volume’, also influenced the amount participants were willing-to-pay for a coffee. If café owners, baristas,

482

and crockery manufacturers want to manipulate people’s expectations of coffee, they should carefully

483

consider the diameter and height of the cups they use/produce, as these features will likely affect expected

484

aroma, bitterness, sweetness, and intensity. Further, these people should be cognizant of traditions (e.g.,

485

serving more concentrated coffees in smaller cups) as they are likely to be important. When providing

486

customers with coffee, café owners and baristas should use a mug shape that conveys a message that is

487

congruent with consumer expectations. This is important because aligning a product with consumer

488

expectations could contribute to product purchasing behaviour. These results add to a growing body of

489

research highlighting the associations between visual information and a product’s likely (or expected)

490

sensory qualities.

491

492 493

REFERENCES

494

Bar, M., & Neta, M. (2006). Humans prefer curved visual objects. Psychological Science, 17(8), 645-648.

495 496

Batra, R., Seifert, C., & Brei, D. (Eds.). (2015). The psychology of design: Creating consumer appeal. London, UK: Routledge.

497 498

Bohrn, I., Nabecker, G., & Carbon, C. C. (2008). Are curved visual objects always preferred? Perception ECVP Abstract, 37, 75-75.

499 500 501

Bremner, A. J., Caparos, S., Davidoff, J., de Fockert, J., Linnell, K. J., & Spence, C. (2013). “Bouba” and “Kiki” in Namibia? A remote culture make similar shape-sound matches, but different shape-taste matches to Westerners. Cognition, 126(2), 165-172.

502 503

Carbon, C. C. (2010). The cycle of preference: Long-term dynamics of aesthetic appreciation. Acta Psychologica, 134(2), 233-244.

504

Cheskin, L. (1957). How to predict what people will buy. New York, NY: Liveright.

505 506

Cliff, M. (2001). The influence of wine glass shape on perceived aroma and colour intensity in wines. Journal of Wine Research, 12, 39-46.

507 508

Deroy, O., & Valentin, D. (2011). Tasting liquid shapes: Investigating the sensory bias of cross-modal correspondences. Chemosensory Perception, 4, 80-90.

509 510

Dijksterhuis, G. (1998). European dimensions of coffee: Rapid inspection of a data set using Q-PCA. Food Quality & Preference, 9, 95-98.

511 512

Gómez-Puerto, G., Munar, E., & Nadal, M. (2015). Preference for curvature: A historical and conceptual framework. Frontiers in Human Neuroscience, 9, 712.

513 514

Hanson-Vaux, G., Crisinel, A. S., & Spence, C. (2013). Smelling shapes: Crossmodal correspondences between odors and shapes. Chemical Senses, 38(2), 161-166.

515 516 517

Harrar, V., & Spence, C. (2013). The taste of cutlery: How the taste of food is affected by the weight, size, shape, and colour of the cutlery used to eat it. Flavour, 2:21, http://www.flavourjournal.com/content/2/1/21

518 519 520

Howie, M. (2012, June 28). We’re tea sick! Survey shows Britain turning to coffee. Retrieved from http://www.standard.co.uk/news/uk/were-tea-sick-survey-shows-britain-turning-to-coffee7895707.html?origin=internalSearch

521 522

Jeon, S.-Y., Lee, E.-K., & Kim, K.-O. (2014). The perceived saltiness of soup affected by tasting protocols. Food Quality and Preference, 35, 98-103.

523 524

Krishna, A., & Morrin, M. (2008). Does touch affect taste? The perceptual transfer of product container haptic cues. Journal of Consumer Research, 34, 807-818.

525 526 527

Lakens, D. (2016, January, 1st). Error control in exploratory ANOVA's: The how and the why [Web blog post]. Retrieved from http://daniellakens.blogspot.co.uk/2016/01/error-control-in-exploratoryanovas-how.html

528 529 530

Ngo, M. K., Velasco, C., Salgado, A., Boehm, E., O’Neill, D., & Spence, C. (2013). Assessing crossmodal correspondences in exotic fruit juices: The case of shape and sound symbolism. Food Quality & Preference, 28(1), 361-369.

531 532

Olson, J. C., & Dover, P. (1976). Effects of expectation creation and disconfirmation on belief elements of cognitive structure. Advances in Consumer Research, 3(1), 168-175.

533 534 535

Piqueras-Fiszman, B., Alcaide, J., Roura, E., & Spence, C. (2012). Is it the plate or is it the food? Assessing the influence of the color (black or white) and shape of the plate on the perception of the food placed on it. Food Quality & Preference, 24(1), 205-208.

536 537

Piqueras-Fiszman, B., Harrar, V., Alcaide, J., & Spence, C. (2011). Does the weight of the dish influence our perception of food? Food Quality & Preference, 22, 753-756.

538 539 540

Piqueras-Fiszman, B., & Spence, C. (2012). The weight of the bottle as a possible extrinsic cue with which to estimate the price (and quality) of the wine? Observed correlations. Food Quality and Preference, 25, 41-45.

541 542 543

Piqueras-Fiszman, B., & Spence, C. (2015). Sensory expectations based on product-extrinsic food cues: An interdisciplinary review of the empirical evidence and theoretical accounts. Food Quality & Preference, 40, 165-179.

544 545

P.

546 547

Prescott, J. (2015). Multisensory processes in flavour perception and their influence on food choice. Current Opinion in Food Science, 3, 47-52.

548 549 550

Salgado-Montejo, A., Alvarado, J. A., Velasco, C., Salgado, C. J., Hasse, K., & Spence, C. (2015). The sweetest thing: The influence of angularity, symmetry, and the number of elements on shapevalence and shape-taste matches. Frontiers in Psychology, 6:1382.

551 552

Shankar, M. U., Levitan, C. A., & Spence, C. (2010). Grape expectations: The role of cognitive influences in color-flavor interactions. Consciousness & Cognition, 19(1), 380-390.

553 554

Shermer, D. Z., & Levitan, C. A. (2014). Red hot: The crossmodal effect of color intensity on perceived piquancy. Multisensory Research, 27, 207-223.

555 556

Silvia, P. J., & Barona, C. M. (2009). Do people prefer curved objects? Angularity, expertise, and aesthetic preference. Empirical Studies of the Arts, 27(1), 25-42.

557

Spence, C. (2011). Crystal clear or gobbletigook? World Fine Wine, 33, 96-101.

558 559

Spence, C. (2012). Managing sensory expectations concerning products and brands: Capitalizing on the potential of sound and shape symbolism. Journal of Consumer Psychology, 22(1), 37-54.

560

Spence, C. (2015a). Multisensory flavor perception. Cell, 161(1), 24-35.

561 562

Spence,

563

Spence, C. (2016). Enhancing the experience through smell. Food Science and Technology, 30(2), 32-35.

564

Spence, C., & Wang, Q. (2015). Sonic expectations: On the sounds of opening and pouring. Flavour, 4, 35.

565 566

Stewart, P. C., & Goss, E. (2013). Plate shape and colour interact to influence taste and quality judgments. Flavour, 2, 27, http://www.flavourjournal.com/content/2/1/27

567 568

Van Doorn, G., Wuillemin, D., & Spence, C. (2014). Does the colour of the mug influence the taste of the coffee? Flavour, 3, 10, http://www.flavourjournal.com/content/3/1/10

569

Velasco, C., Woods, A. T., Hyndman, S., & Spence, C. (2015). The taste of typeface. i-Perception, 6, 1-10.

570 571 572

Velasco, C., Woods, A. T., Petit, O., Cheok, A. D., & Spence, C. (2016). Crossmodal correspondences between taste and shape, and their implications for product packaging: A review. Food Quality and Preference, 52, 17-26.

573 574 575

Wan, X., Woods, A. T., van den Bosch, J., Mckenzie, K. J., Velasco, C., & Spence, C. (2014). Crosscultural differences in crossmodal correspondences between tastes and visual features. Frontiers in Psychology: Cognition, 5, 1365.

576 577

Wansink, B., & Van Ittersum, K. (2003). Bottoms up! The influence of elongation on pouring and consumption volume. Journal of Consumer Research, 30(3), 455-463.

578 579

Wansink, B., & van Ittersum, K. (2005). Shape of glass and amount of alcohol poured: Comparative study of effect of practice and concentration. British Medical Journal, 331, 1512-1514.

580 581 582

Westerman, S. J., Gardner, P. H., Sutherland, E. J., White, T., Jordan, K., Watts, D., & Wells, S. (2012). Product design: Preference for rounded versus angular design elements. Psychology & Marketing, 29(8), 595-605.

J.

W., & D. H. (2013, December 16). The coffee insurgency. http://www.economist.com/blogs/graphicdetail/2013/12/daily-chart-17

Retrieved

C. (2015b). On the psychological impact of food colour. Flavour, https://flavourjournal.biomedcentral.com/articles/10.1186/s13411-015-0031-3

4,

from

21,

583 584

Williams, L. E., & Bargh, J. A. (2008). Experiencing physical warmth promotes interpersonal warmth. Science, 322(5901), 606-607.

585 586 587

Yeomans, M. R., Chambers, L., Blumenthal, H., & Blake, A. (2008). The role of expectancy in sensory and hedonic evaluation: The case of smoked salmon ice-cream. Food Quality & Preference, 19(6), 565-573.

588 589

Zhang, Y., Feick, L., & Price, L. J. (2006). The impact of self-construal on aesthetic preference for angular versus rounded shapes. Personality and Social Psychology Bulletin, 32(6), 794-805.

590 591

Appendix

592 593 594 595 596

Table 3: The results of 8 separate mixed-factorial ANOVAs, one for each of the dependent variables. As these were exploratory analyses, the Holm-Bonferroni multiple-comparison correction incorporated both the number of dependent variables and the number of separate comparisons for each ANOVA (maximum critical alpha was thus 0.05 / 8 x 15 = 0.00042, see Lakens, 2016). Significant factors and interactions less than this critical alpha have been suffixed with a plus-sign.

Critical alpha Factors

Aroma

Bitter

df

F

Sig.

Partial Eta Squared

Country of origin

2

4.230

0.015

0.001

0.027

Cup diameter

1

13.778

0.000

0.000

0.044 +

Thickness of rim

1

2.447

0.119

0.001

0.008

Height of cup

1

45.734

0.000

0.000

0.132 +

Diameter * Country

2

2.110

0.123

0.001

0.014

Thickness * Country

2

0.245

0.783

0.005

0.002

Height * Country

2

0.859

0.425

0.001

0.006

Diameter * Thickness

1

0.420

0.518

0.002

0.001

Diameter * Thickness * Country

2

0.146

0.864

0.007

0.001

Diameter * Height

1

0.146

0.703

0.004

0.000

Diameter * Height * Country

2

0.777

0.461

0.001

0.005

Thickness * Height

1

0.041

0.840

0.006

0.000

Thickness * Height * Country

2

1.667

0.191

0.001

0.011

Diameter * Thickness * Height Diameter * Thickness * Height * Country

1

3.804

0.052

0.001

0.012

2

0.206

0.814

0.005

0.001

Country of origin

2

2.065

0.129

0.001

0.013

Cup diameter

1

137.560

0.000

0.000

0.313 +

Thickness of rim

1

0.537

0.464

0.001

0.002

Height of cup

1

69.037

0.000

0.000

0.186 +

Diameter * Country

2

1.414

0.245

0.001

0.009

Thickness * Country

2

5.012

0.007

0.001

0.032

Height * Country

2

0.011

0.989

0.050

0.000

Diameter * Thickness

1

0.045

0.833

0.006

0.000

Diameter * Thickness * Country

2

0.029

0.971

0.025

0.000

Energy

Temp.

Diameter * Height

1

3.250

0.072

0.001

0.011

Diameter * Height * Country

2

1.991

0.138

0.001

0.013

Thickness * Height

1

0.019

0.891

0.008

0.000

Thickness * Height * Country

2

9.317

0.000

0.000

0.058 +

Diameter * Thickness * Height Diameter * Thickness * Height * Country

1

1.993

0.159

0.001

0.007

2

1.274

0.281

0.001

0.008

Country of origin

2

7.421

0.001

0.000

0.047

Cup diameter

1

5.521

0.019

0.001

0.018

Thickness of rim

1

0.294

0.588

0.002

0.001

Height of cup

1

3.831

0.051

0.001

0.013

Diameter * Country

2

3.264

0.040

0.001

0.021

Thickness * Country

2

0.355

0.701

0.003

0.002

Height * Country

2

0.826

0.439

0.001

0.005

Diameter * Thickness

1

0.006

0.937

0.017

0.000

Diameter * Thickness * Country

2

2.571

0.078

0.001

0.017

Diameter * Height

1

11.905

0.001

0.000

0.038

Diameter * Height * Country

2

5.240

0.006

0.001

0.034

Thickness * Height

1

0.507

0.477

0.001

0.002

Thickness * Height * Country

2

0.364

0.695

0.003

0.002

Diameter * Thickness * Height Diameter * Thickness * Height * Country

1

0.173

0.678

0.003

0.001

2

1.102

0.334

0.001

0.007

Country of origin

2

12.893

0.000

0.000

0.079 +

Cup diameter

1

5.711

0.017

0.001

0.019

Thickness of rim

1

0.159

0.690

0.003

0.001

Height of cup

1

0.897

0.344

0.001

0.003

Diameter * Country

2

0.261

0.771

0.004

0.002

Thickness * Country

2

0.361

0.697

0.003

0.002

Height * Country

2

2.866

0.058

0.001

0.019

Diameter * Thickness

1

0.015

0.903

0.010

0.000

Diameter * Thickness * Country

2

0.943

0.390

0.001

0.006

Diameter * Height

1

1.507

0.221

0.001

0.005

Diameter * Height * Country

2

5.301

0.005

0.001

0.034

Thickness * Height

1

1.470

0.226

0.001

0.005

Thickness * Height * Country

2

1.296

0.275

0.001

0.009

Diameter * Thickness * Height Diameter * Thickness * Height * Country

1

0.441

0.507

0.002

0.001

2

1.420

0.243

0.001

0.009

2

6.369

0.002

0.000

0.040

Cup diameter

1

110.671

0.000

0.000

0.268 +

Thickness of rim

1

6.276

0.013

0.001

0.020

Intensity Country of origin

Liking

Money

Height of cup

1

81.507

0.000

0.000

0.213 +

Diameter * Country

2

2.987

0.052

0.001

0.019

Thickness * Country

2

0.699

0.498

0.002

0.005

Height * Country

2

0.742

0.477

0.001

0.005

Diameter * Thickness

1

4.662

0.032

0.001

0.015

Diameter * Thickness * Country

2

0.914

0.402

0.001

0.006

Diameter * Height

1

2.589

0.109

0.001

0.008

Diameter * Height * Country

2

3.966

0.020

0.001

0.026

Thickness * Height

1

4.021

0.046

0.001

0.013

Thickness * Height * Country

2

0.996

0.370

0.001

0.007

Diameter * Thickness * Height Diameter * Thickness * Height * Country

1

0.287

0.593

0.002

0.001

2

1.281

0.279

0.001

0.008

Country of origin

2

2.900

0.057

0.001

0.019

Cup diameter

1

6.078

0.014

0.001

0.020

Thickness of rim

1

2.178

0.141

0.001

0.007

Height of cup

1

11.844

0.001

0.000

0.038

Diameter * Country

2

5.335

0.005

0.001

0.034

Thickness * Country

2

0.683

0.506

0.002

0.005

Height * Country

2

9.896

0.000

0.000

0.062 +

Diameter * Thickness

1

0.207

0.649

0.002

0.001

Diameter * Thickness * Country

2

0.393

0.675

0.003

0.003

Diameter * Height

1

1.587

0.209

0.001

0.005

Diameter * Height * Country

2

0.516

0.598

0.002

0.003

Thickness * Height

1

1.495

0.222

0.001

0.005

Thickness * Height * Country

2

0.919

0.400

0.001

0.006

Diameter * Thickness * Height Diameter * Thickness * Height * Country

1

1.863

0.173

0.001

0.006

2

0.662

0.517

0.002

0.004

Country of origin

2

6.963

0.001

0.000

0.044

Cup diameter

1

90.621

0.000

0.000

0.231 +

Thickness of rim

1

0.274

0.601

0.002

0.001

Height of cup

1

66.102

0.000

0.000

0.180 +

Diameter * Country

2

28.706

0.000

0.000

0.160 +

Thickness * Country

2

1.326

0.267

0.001

0.009

Height * Country

2

20.040

0.000

0.000

0.117 +

Diameter * Thickness

1

0.707

0.401

0.001

0.002

Diameter * Thickness * Country

2

0.132

0.877

0.007

0.001

Diameter * Height

1

12.828

0.000

0.000

0.041 +

Diameter * Height * Country

2

2.620

0.074

0.001

0.017

Thickness * Height

1

2.317

0.129

0.001

0.008

Thickness * Height * Country

2

0.668

0.514

0.002

0.004

Sweet

597 598

Diameter * Thickness * Height Diameter * Thickness * Height * Country

1

5.390

0.021

0.001

0.018

2

1.859

0.158

0.001

0.012

Country of origin

2

6.348

0.002

0.001

0.040

Cup diameter

1

33.552

0.000

0.000

0.100 +

Thickness of rim

1

0.470

0.493

0.002

0.002

Height of cup

1

2.457

0.118

0.001

0.008

Diameter * Country

2

6.715

0.001

0.000

0.043

Thickness * Country

2

0.551

0.577

0.002

0.004

Height * Country

2

4.568

0.011

0.001

0.029

Diameter * Thickness

1

2.325

0.128

0.001

0.008

Diameter * Thickness * Country

2

1.985

0.139

0.001

0.013

Diameter * Height

1

7.687

0.006

0.001

0.025

Diameter * Height * Country

2

4.289

0.015

0.001

0.028

Thickness * Height

1

3.707

0.055

0.001

0.012

Thickness * Height * Country

2

0.095

0.910

0.013

0.001

Diameter * Thickness * Height Diameter * Thickness * Height * Country

1

2.380

0.124

0.001

0.008

2

0.755

0.471

0.001

0.005

599 600 601 602 603 604 605

Highlights • • • •

Shape-taste expectations elicited by pictures of mugs were examined. The relevant research about crossmodal associations is highlighted and reviewed. The width and height of the mugs was shown to be important. Findings highlight the complex nature of shape-flavour interactions.